espite ambiguity and shifting priorities from the US FDA, most sponsors remain committed to improving diversity in clinical trials not as a compliance exercise, but as a scientific and operational imperative.

In the context of the current state of diversity in clinical research, sponsors are putting intent into practice through earlier feasibility planning, more thoughtful site selection, and patient-centric trial design, while balancing scientific goals with practical considerations such as cost, resourcing, and timelines. Drawing on real-world data and industry experience, this highlights how diversity considerations are integrated earlier and more consistently across the trial lifecycle, even amid a less prescriptive regulatory environment.

Between Statutory and Discretionary: DAPs Remain in Limbo

June 2024 saw the FDA release its most recent draft guidance on Diversity Action Plans (DAPs). This recommends that studies specify their goals for enrollment, disaggregated by race, ethnicity, sex, and age. While the requirement for DAPs has been established through the Food and Drug Omnibus Reform Act (FDORA), its implementation remains contingent on final FDA guidance. Sponsors will have a 180-day implementation period following publication of this guidance, the timing of which remains uncertain.

The intervening period has seen the draft diversity guidance removed and later restored to the FDA’s website as regulatory and legislative priorities have shifted. In the year prior to September 2024, FDA’s CDER received 161 DAPs, a 30% year-on-year increase, but no updated reporting is available since then. According to data from Citeline, more than 1,000 mid- to late-stage trials begin each year in the US. This clearly leaves a large gap between the present reality and an aspirational target.

Pharmaceutical and Biotech Companies Face Different Challenges

Clinical trial diversity cannot ignore a different kind of diversity: the one across trial sponsors.

The challenges facing a top 10 pharmaceutical company are fundamentally different from those facing a Series B biotech. Yet, diversity expectations, whether explicit or implicit, apply to both. And both must adopt a pragmatic approach that works for them.

For large pharmaceutical companies, DAPs have largely been absorbed into standard operating procedures. Generally, these companies have the budgets, knowledge, and data infrastructure to employ a diversity-first approach from the outset. Additionally, many have public diversity commitments and are tracking enrollment demographics as a standard metric.

However, we have observed one large pharmaceutical company tripped up by the current regulatory ambiguity. Despite crafting a comprehensive DAP for a pivotal phase 3 trial, the resulting patient population was arbitrarily narrow. As a consequence, approval was granted but with a restricted product label, which will cost hundreds of millions of dollars to expand with new trials. Had they adopted a more encompassing approach to diversity, this added cost and delay to market could have been avoided.

On the flip side, many biotechs have reservations about the practicalities that diversity in trials can bring. Building out community engagement infrastructure, qualifying decentralized sites, and procuring the necessary data are not trivial expenses for organizations in a constrained funding environment. And considering that many biotech programs are targeting rare diseases and niche populations, there are valid concerns about how diversity obligations will affect time to market. With a finite funding runway, biotechs are obliged to find the quickest path to market, both for the patients they wish to serve and their shareholders.

AI and RWD Enabling the Solutions

The business, scientific, and ethical case for a pragmatic, scientifically grounded approach to diversity in clinical trials is not new. However, the lull in US regulatory updates affords to trial sponsors the opportunity to build the right foundations and practical frameworks for their diversity ambitions.

Thankfully, an AI- and real-world data-led revolution is arriving at just the right time to help trial sponsors balance diversity with pragmatism. We are seeing a lot of appetite from across the clinical spectrum for how these tools can be used to better understand a range of features that tie into DAPs:

• Standard of care and unmet needs
• Patient demographics and social determinants of health
• Feasibility modeling
• Site and investigator selection
• Protocol design
• Patient recruitment and engagement.

Our experience in developing these tools and supporting trial sponsors in diversity initiatives distills into two core recommendations.

Start with feasibility, not aspiration

One of the most consistent findings from sponsors who have successfully improved trial diversity is that the work begins before protocol finalization. Feasibility assessments that incorporate demographic data on likely patient pools by indication, geography, and existing site networks allow sponsors to set realistic enrollment goals and identify gaps before they become midtrial problems.

This means integrating patient-availability data into decision-making. One sponsor recently increased their predicted enrollment rate by 20% by diagnosing inclusion and exclusion constraints and proposing protocol amendments to expand their patient pool. This early catch not only saved the sponsor time and resources but also broadened their eligible participant base.

Meet patients where they are

Even after a trial is operational, study sponsors can still hugely influence diversity outcomes. Every day, billions of healthcare data points are added to electronic health records, lab results, and prescription claims. The latency in such real-world data has reduced from monthly updates to live querying such that a near-real-time picture of a treatment landscape can emerge. This allows clinical teams to identify and match potentially eligible patients to trials faster than ever before.

As this identification occurs increasingly digitally, many of the systemic biases in the healthcare system can be bypassed. In particular, AI can hugely enrich data sets like unstructured medical notes, which contain highly relevant patient history that can be matched against inclusion and exclusion criteria for studies. Once a new patient emerges, targeted physician outreach and referrals can dramatically boost enrollment, all within a tight window of eligibility. In one example, real-world data triggers drove 64% of randomized patients across 25 sites within a 4-month recruitment period, accelerating timelines by 2 months. This real-time matching reaches patients that would otherwise go under the radar, reducing screen fail rates and lowering site burden.

Closing Thoughts

There remains a strong scientific and moral consensus that the pharmaceutical industry needs to move forward with diversity, both in clinical trials and in patient care. This is occurring independently of regulation, and we advise sponsors not to predict or pre-empt the timelines for enforcement. However, the difference between the FDA’s disclosed DAPs and current late-stage trial activity suggests that there is a significant gap to bridge.

Thankfully, new and proven solutions are on hand to support trial sponsors in achieving a balance between the scientific challenges and commercial practicalities that trial diversity poses. This pragmatism needn’t come at the expense of availability of new treatments. If industry can get this right, everyone can benefit—including underserved patients at the heart of this initiative.